The antitumor activity of human T cells, as definitively demonstrated by donor lymphocyte infusions following allogeneic hematopoietic stem cell transplantation (HSCT), has provided impetus for developing T cell therapy as a modality to treat leukemia. However, identifying immunogenic antigens expressed by leukemic cells has been a major obstacle. Self-proteins overexpressed by cancer cells have been shown in some settings to induce T cell responses with avidity sufficient to recognize transformed cells but not normal cells which express lower levels of the protein. The best candidate self-proteins are the subset which are oncogenic and/or associated with maintenance of the malignant phenotype, as these may be indispensable to the leukemia. Two proteins, WT1 and Proteinase 3/myeloblastin (PR3/MBN), expressed by acute leukemias and CML respectively, have recently been demonstrated in animal models, and preclinical and clinical studies, to potentially fulfill criteria for oncogenicity, immunogenicity, and safety as targets. Our laboratory has extensive experience with adoptive transfer of cloned T cells, which can facilitate analysis of candidate antigens by establishing in vivo qualitative and quantitative responses not readily achieved by other means. This approach can not only help define the therapeutic potential of specifically targeting a selected antigen, but determine the possible risks and toxicities to normal tissues, and the magnitude of response that would need to be achieved by alternative approaches such as vaccination. The proposed experiments are designed to evaluate the immunogenicity of the WT1 and PR3/MBN proteins, and to determine if adoptive transfer of CD8+ CTL clones specific for these antigens can mediate an antileukemic effect without toxicity to the recipient. The specific aims are to: 1. Determine if CD8+ T cells specific for WT1 and PR3/MBN and capable of selectively recognizing leukemic cells can be isolated from an MHC diverse population of normal individuals and patients with leukemia. 2. Determine if expanded numbers of CD8+ CTL specific for immunogenic epitopes from WT1 and PR3/MBN are present in leukemic patients prior to and following HSCT. 3. Determine if the transfer of donor-derived CD8+ CTL specific for PR3/MBN into patients with CML-AP or CML-BP that has relapsed post-allogeneic HSCT is safe and can mediate an antitumor effect. 4. Determine if the transfer of donor-derived CD8+ CTL specific for WT1 into patients with AML or ALL that has relapsed post-allogeneic HSCT is safe and can mediate an antitumor effect.